Energy Conversion and Management 48 (2007) 155–165

www.elsevier.com/locate/enconman

Mathematical modeling and numerical simulation of a short

term thermal energy storage system using phase change
material for heating applications
Miroslaw Zukowski *

Department of Heat Engineering, Bialystok Technical University, Wiejska 45A, 15-351 Bialystok, Poland

Received 28 October 2005; accepted 28 April 2006

Available online 19 June 2006

Abstract

The main objective of this paper is to analyze the heat and mass transfer in a ventilation duct filled with encapsulated paraffin wax
RII-56. A mathematical model and its numerical solution are presented. A new approach for approximating the specific heat of the PCM
as a function of its temperature is proposed for all range of operating conditions. The interpolating cubic spline function method is used
for determining an effective specific heat in each time step of temperature calculations. Equations for three-dimensional transient thermal
analysis with associated boundary and initial conditions are solved by the control volume finite difference method with the fully implicit
scheme. Proper trends and a good agreement between the results of theoretical modeling and physical reality for the solidification/melt-
ing process are obtained. Finally, 24 h operation of the proposed short term latent heat storage system is analyzed and the influence of
the PCM capsules geometry on the thermal performance of the tested storage module is shown.
 2006 Elsevier Ltd. All rights reserved.

Keywords: Thermal energy storage (TES); Phase change material (PCM); Effective heat capacity method; Control volume finite difference method

1. Introduction different heat exchanger constructions, is presented by

Zelba et al. [1]. Investigations of TES systems incorporat-
Thermal energy storage (TES) applications have signifi- ing PCMs (phase change material) for use in building
cantly increased because of changes in energy price and applications are widely summarized in the Khudhair and
changes in environmental regulation. TES units can work Farid review [2]. Only these researches that refer to heat
as a long term or a short term (few hours to several days) transfer geometries that are similar to the tested unit, were
storage device. The second type is more popular for use in mentioned in the current paper.
HVAC systems designed for single family houses, and the The Second Law analysis was used by Watanabe and
current paper is focused on this technology. Kanzawa [3] to determine the optimum melting point dis-
Modeling of heat transfer in melting and solidification tribution of PCMs. Their model described a heat storage
processes has been spread widely in the scientific literature unit that consisted of an encapsulated PCM in horizontal
for the last 30 years. A very good review on thermal energy cylindrical tubes [4]. The capsules were fixed in an in line
storage, particularly on moving boundary problems, a arrangement. The energy equation for water flowing across
numerical solution for conduction only, models both for the capsules was formulated in unsteady state conditions
conduction and convection and numerical simulations in and for one-dimensional geometry. The fluid temperature
was calculated from the finite difference form and explicit
method. The authors concluded that the melting point dis-
*
Tel.: +48 85 7469680. tribution of the PCMs has a substantial effect on the charg-
E-mail addresses: [email protected], [email protected]. ing and discharging rates as well as the exergy efficiency.

0196-8904/$ - see front matter  2006 Elsevier Ltd. All rights reserved.
doi:10.1016/j.enconman.2006.04.017
156 M. Zukowski / Energy Conversion and Management 48 (2007) 155–165

Nomenclature

A surface area (m2) dt duration of day tariff (h)

C equivalent thermal conductivity coefficient q density (kg/m3)
c specific heat (J/kg K)
d distance between PCM capsules (m) Subscripts
h convective heat transfer coefficient (W/m2 K) I day tariff
E thermal energy (J) A air
H enthalpy (J/kg) CV control volume
k thermal conductivity (W/mK) eq equivalent
l PCM thickness (m) D demand
m, n equivalent thermal conductivity exponent G support construction
Q heat transfer rate (W) IN inlet
Ra Rayleigh number m measurement
S volumetric rate of heat generation (W/m3) OUT outlet
t time (s) P, PCM phase change material
T temperature (K) p constant pressure
w velocity (m/s) p predicted
V volume (m3) SF surface
V volume rate of flow (m3/s) W wall
x, y, z coordinates (m) 0 starting value
1 ending value
Greek symbols
d PCM liquid thickness (m)
d denotes difference

Arkar and Medved [5] described a short time storage proved that the use of polyethylene or PVC for the spher-
module that can be a heat exchanger and ventilation duct ical capsule facilitates the construction of the storage and
at the same time. The investigated unit was based on poly- reduces its costs.
ethylene spheres with an encapsulated PCM. A one-dimen- Numerical analysis of a PCM bed storage system was
sional numerical model described the heat transfer inside realized by Pluta. In a mathematical model, presented in
the duct with porous media. It was assumed that the melt- Ref. [7], the storage unit was split into separate segments.
ing and solidification processes occur at the same constant An energy balance for the air and the storage material
temperature, and in the liquid phase, convection was not was calculated inside every section. The author concluded
considered. The storage was divided into n identical parts, that a storage bed composed from separate elements has
and the outlet air temperature from each segment was thermal characteristics very close to those of the PCM
determined with the NTU correlation for a typical heat alone storage.
exchanger. The model was verified with measurements. The main goals of the author’s project are to determine
Relatively large deviations between the simulations and the charge and discharge characteristics of a TES prototype
the measured data were obtained during the melting pro- device and to create a computational model for designing
cess. At solidification, the accuracy was good. Measured this module filled with an encapsulated PCM. The analyzed
and calculated results showed that it is possible to prevent TES unit consists of the following components: a Plexiglas
or to reduce building overheating by using the short term casing, light steel nets and PCM enclosed in polyethylene
heat storage unit tested by the authors. film bags. The support construction is used to keep a uni-
A storage unit composed of spherical capsules filled with form distance between the PCM layers. The paraffin wax
PCM placed inside a cylindrical tank was investigated both RII-56 is selected as a storage material. The tested unit is
numerically and experimentally by Ismail and Henriquez designed to cooperate with a warm air based heating sys-
[6]. A numerical model to simulate the characteristics of tem widely described, among others in Refs. [8,9], as a part
the system was based on the tank division into a number of ventilation ducts. Fig. 1 presents an idea of an example
of axial layers. It was assumed that the temperature of construction of this device. The first stage of the project
the fluid was uniform and equal to the average temperature was realized and the effects of the experimental investiga-
of the layer. The solidification process was treated by using tions are reported in Ref. [10]. The numerical simulations,
only one-dimensional heat conduction. The finite difference which are the results of the second stage of the project, are
approximation and a moving grid inside the spherical cap- presented in the current paper. The literature review shows
sules were used to solve the differential equations. They that the most popular methods used to model the phase
 M. Zukowski / Energy Conversion and Management 48 (2007) 155–165 157

Fig. 1. Schematic of the heat storage module.

change process are the enthalpy method and the effective For constant thermophysical properties of the PCM, we
heat capacity method. The last scheme is applied in the can reduce the above relation to:
present investigation. Most commonly, the heat capacity oH P
of the PCM for each temperature interval is given by a con- qP ¼ k P r2 T P þ S: ð2Þ
ot
stant value or is approximated by a polynomial function.    
The author proposes another solution based on an interpo- If the term oHotP is replaced with the term cp;P oTotP and
lation of measurements by cubic spline functions. This S = 0 in Eq. (2), the final form of the energy conservation
solution enables the approximation of each curve (even equation can be obtained
complicated) with a very good accuracy. oT
qP cp;P ¼ k P r2 T P : ð3Þ
ot
2. Mathematical model
In this analysis, the effective heat capacity method to de-
scribe phase change phenomena is adopted. This approxi-
2.1. Assumptions of the mathematical model
mation assumes that specific heat of the PCM depends on
temperature cp,P = f(TP). The relation between specific heat
The short term TES unit considered in the present inves-
and temperature used in the present investigation and the re-
tigation is shown in Fig 1, and the following assumptions
sults of the numerical interpolation are shown in Fig. 2.
have been made for the mathematical model:
The author proposes the interpolating cubic spline func-
tion method (widely described in Refs. [11–13]) for deter-
• The PCM is homogenous and isotropic.
mining the value of cp,P in each time step of the
• Forced convection for the air flow domain is analyzed.
temperature calculations. In this method, it is necessary
• The radiation heat transfer inside the unit was neglected
to declare local values of the material specific heat for the
(very small difference between surface temperatures of
characteristic points of the curve (change of the trend)
PCM and inner walls).
obtained from measurements.
• Heat flux is isotropic on the whole PCM capsule area.
Two stages can be separated in the algorithm of the
• The thermal conductivity and density of PCM are differ-
spline interpolation. At first, we must solve the following
ent for the solid and liquid phases.
linear system of equations:
• Thermal capacity of the unit’s casing is neglected.

2.2. Energy conservation equations

2.2.1. Phase change material

The energy equation (1) for a material undergoing a
phase transformation is given by the following form:
oH P
qP wrH P ¼ k P r2 T P þ S:
þ qP ~ ð1Þ ð4Þ
ot
158 M. Zukowski / Energy Conversion and Management 48 (2007) 155–165

45000 The natural convection effect in the melt region is simu-

40000 lated through an equivalent thermal conductivity keq,P gi-
Cp,p=f(Tp)
Spline interpolation ven by the following correlation:
Specific heat [J/(kgK)]

35000
 m
30000 k eq;P d
¼ CRan : ð6Þ
25000 kP l
20000
The coefficients in Eq. (6) are found to be C = 0.05,
15000 n = 0.25 and m = 1. This kind of estimation has been suc-
10000 cessfully used in several works, among others in Refs. [14–
17].
5000

0
35 37 39 41 43 45 47 49 51 53 55 57 59 61 2.2.2. Air inside the TES unit
TP [ºC] The energy balance for the working fluid (air) on a sub-
domain n of the TES unit, shown in Fig. 3, leads to the fol-
Fig. 2. Apparent specific heat of the paraffin wax, adopted in simulations, lowing equation:
as a function of temperature.
V A;n oT A
qA cp;A
where d j ¼ T jp;P  T j1
p;P ðj ¼ 0; 1; 2; . . . ; n  1Þ, n + 1 is the V CV;n ot
number of division points, ¼ qA cp;A wA;x ðT A;nðeÞ  T A;nðnÞ Þ þ ny nz bP ðT P;SF  T A;n Þ
d jþ1 þ bG ðT G;n  T A;n Þ þ bW ðT A;n  T W;n Þ; ð7Þ
aj ¼ ;
d j þ d jþ1
dj where
bj ¼ ;
d j þ d jþ1
! V CV;n ¼ H y H z dx;
6 cjþ1 j
p;P  cp;P cjp;P  cj1
p;P
cj ¼  : AP;n
d j þ d jþ1 d jþ1 dj bP ¼ hP ;
V CV;n
At the second step, the specific heat of PCM for AG;n
bG ¼ hG ;
T P 2 ½T jP ; T jþ1
P  can be calculated from: V CV;n
cp;P ðT P Þ ¼ cjp;P þ ej ðT P  T jP Þ þ fj ðT P  T jP Þ þ gj ðT P  T jP Þ ;
2 3 A W;n
bW ¼ hW ;
V CV;n
ð5Þ
AW;n ¼ ð2H y þ 2H z Þ dx:
where
cjþ1 j
p;P  cp;P 2M j þ M jþ1 2.2.3. Support construction
ej ¼  d jþ1 ;
d jþ1 6 Conservation of energy for the light steel nets that sup-
Mj port the PCM capsules is given by the following equation:
fj ¼ ;
2
M jþ1 þ M j V G oT G
gj ¼ : qG cG ¼ bG ðT A;n  T G;n Þ: ð8Þ
6d jþ1 V A;n ot

Fig. 3. Three-dimensional domain of the model (divided into several sub-domains).

 M. Zukowski / Energy Conversion and Management 48 (2007) 155–165 159

2.3. Initial and boundary conditions Table 1

The system of linear equations that describes heat and mass transfer in the
TES unit
The constant temperature field in the TES module is
assumed as an initial condition AT B
9 The balance of energy for the PCM
0 6 x 6 Hx > aP T tþdt tþdt tþdt tþdt
qP cp;P DxDyDzT tP;P
= P;P  aE T P;E  aW T P;W  aN T P;N
t¼0
0 6 y 6 Hy ! T A ¼ T t¼0 t¼0 t¼0
P ¼ T G ¼ T W ¼ T 0: ð9Þ tþdt tþdt tþdt
aS T P;S  aT T P;T  aB T P;B
>
; where
0 6 z 6 Hz aP = aP1  aE  aW  aN  aS  aT  aB,
aP1 = qPcp,PDxDyDz,
Boundary conditions: kP
aE ¼ ðdxÞ DyDzdt,
e
kP
aW ¼ ðdxÞ DyDzdt,
air–PCM interface kP
w

aN ¼ ðdxÞ DxDzdt,
n
oT P kP
aS ¼ ðdxÞ DxDzdt,
k P ¼ hP ðT A  T P;SF Þ; ð10Þ s
on kP
aT ¼ ðdxÞ DxDydt,
t
air–wall interface (inside the unit) kP
aB ¼ ðdxÞ DxDydt.
b

oT W In the liquid phase kP is replaced with keq,P

k W ¼ hW ðT A  T W;SFðiÞ Þ; ð11Þ
on The balance of energy for the working fluid on sub-domain n
aP T tþdt tþdt tþdt tþdt
A;n  aV T A;n1  aV T A;nþ1  aBP T P;SF qA cp;A V A;n T tA;n ,
air–wall interface (outside the unit)
aCW T tþdt tþdt
W;n  aGA T G;n
oT W where
k W ¼ hOUT ðT A  T W;SFðeÞ Þ; ð12Þ aP = qAcp,AVA,n + aBP + aCW + aGA,
on
aV = qAcp,AwA,xDyDzdt,
inlet plane aBP = nynzbPVCV,ndt,
aGA = qGcGVG,n.
0 6 y 6 H y ! wA;x ¼ wA;IN ; T A ¼ T A;IN ; ð13Þ
The balance of energy for the support construction on sub-domain n
0 6 z 6 H z ! wA;y ¼ wA;z ¼ 0; ð14Þ tþdt
aP T G;n  aGB T tþdt
A;n qG cG V G;n T tG;n ,
outlet plane where
 aP = qAcp,AVA,n + aBP + aCW + aGA,
0 6 y 6 Hy owA;x oT A aV = qAcp,AwA,xDyDzdt,
! ¼ 0; ¼ 0; ð15Þ aBP = nynzbVVCV,ndt,
0 6 z 6 Hz ox ox
aGA = qGcGVG,n.
0 6 z 6 H z ! wA;y ¼ wA;z ¼ 0: ð16Þ

3. Numerical solution
T
N
The control volume finite difference method (CVFDM)
of Patankar [18] is used to solve Eq. (3). A fully implicit
t
method (with r = 1), also called the backward Euler (δz)t (δy)n
scheme, defined by relation (17) is chosen for the tempera- n
ture discretization in time E
W w P e
Z tþdt (δx)w (δx)e

T P dt ¼ ½rT Ptþdt þ ð1  rÞT tP dt ð17Þ Δz s

t (δy)s b Δy
(δz)b
Integrating the energy equations over the control volumes
and time leads to the finite difference scheme. The resulting S
Δx
system of linear equations AT = B is solved using the stan- B
dard Gauss elimination method. Table 1 shows the coeffi-
cient vectors A and B. Moreover, the particular symbols Fig. 4. An internal control volume of the PCM domain.
used in Table 1 are presented in Fig. 4.
The boundary condition relations that complete the sys-
tem of linear equations are not presented in Table 1 shape of a single polyethylene bag. The dimensions of the
because of their large quantity. tested capsules are the following: width 31.5 mm, height
Because of the phase transformation, iteration of calcu- 16.0 mm and length 31.5 mm. The mathematical model,
lations is used during each time step. The convergence cri- and its numerical solution presented above, are imple-
terion for temperature is set to 0.001 K. The calculations mented as a computer program TES_MZ, which simulates
are executed only for one eighth of the PCM capsule (see the heat transfer characteristic of the tested unit. Valida-
Fig. 5). In this study, a rectangular prism approaches the tion of this code is presented in the next chapter.
160 M. Zukowski / Energy Conversion and Management 48 (2007) 155–165

q = hP (TA–TP,SF) 200
q=0

150 VA,IN = 20 m3/h

100

50

0
q=0 250
Experiment
200
Simulation
Fig. 5. Dividing the PCM capsule model into eight parts.

QPCM [W]
150

100 VA,IN = 30 m3/h

4. Validation test of the model
50
To estimate the accuracy of the numerical simulations,
0
the experimental data presented in Ref. [10] is compared
300
with results of calculations. The thermophysical properties
of the tested paraffin wax RII-56 were prepared on the 250 VA,IN = 40 m3/h
basis of the measurement and manufacturers’ data. They 200
are summarized and presented in Table 2.
150
Calculations were performed for different value of the
inlet air temperature (60, 65 and 70 C for charging; 100
21 C for discharging). The ambient air temperature is set 50
to 21 C for all cases. A comparison between the measured
0
and predicted values of the heat flux QPCM (the heat flux is 0 25 50 75 100 125 150 175 200
absorbed by the PCM during the charging period, see t [min]
Fig. 6, and emitted during the cooling process, see Fig. 7)
Fig. 6. The heat flux absorbed by the PCM as a function of time.
is presented only for one selected inlet air temperature
TA,IN = 65 C, but for three air flow rates (20, 30 and
The relative accuracy in the final charging and discharg-
40 m3/h).
ing times was less than 9% and 14%, respectively. From
The relative maximum error, given by Eq. (18), has been
these validation test results, it can be concluded that the
adopted to quantify the difference between the experimen-
theoretical model gives acceptable accuracy, and the com-
tal data and the calculation results
   puter code TES_MZ can be used successfully to predict
xm  xp  the thermal performance of the tested storage device.
RME ¼ max     100% : ð18Þ
i¼1;2;...;n xp 
5. Results of calculations
The value of RME in the heat flux absorbed by the paraffin
wax is less than 12%. For the discharging process, this va-
The influence of the distance between PCM capsules on
lue is less than 10%. As Figs. 6 and 7 show, the agreement
the thermal performance of the storage system is analyzed.
between physical reality and numerical simulations is rea-
The assumptions for the calculations, based on the com-
sonable for the melting as well as for the solidification pro-
puter code TES_MZ, are presented in Table 3.
cess. However, it can be observed that the theoretical
As Fig. 10 shows, the final charging and discharging
model predicts a little shorter charging time for the tested
time depends exponentially on the capsules density, but
device.
the maximum difference is quite small and is less than
Furthermore, the measured and computed final charg-
15–18%. If we decrease the distance between capsules, the
ing (Fig. 8) and discharging time periods (Fig. 9) are
TES device volume is reduced, but on the other hand, the
compared.
noise level and the pressure losses increase considerably.
It is hard to indicate the optimal solution for the designer.
Nevertheless, we should pursue solutions to minimize the
volume of the storage unit.
Table 2
Thermophysical properties of the paraffin wax RII-56 Also, Figs. 11–13 indicate the influence of the capsule
volume on the TES device thermal characteristics.
qPCM cPCM kPCM HPCM Temperature range of fusion
A logarithmic trend line is a best fit to the calculation
842 2174 0.267 156 49–57
results presented in Fig. 11. In this case, a great difference
 M. Zukowski / Energy Conversion and Management 48 (2007) 155–165 161

200 250
225
150 VA,IN = 20 m3/h 200
175
150 TPCM,0 -TPCM, ∞ = 39 ºC
100
125
180
50 Simulation
160

t [min]
Experiment
140
0 120
250 100 TPCM,0 -TPCM, ∞ = 44 ºC
Experiment 80
200 130
Simulation
QPCM [W]

115
150
100
100 VA,IN = 30 m3/h 85
TPCM,0 -TPCM, ∞ = 49 ºC
50 70
20 25 30 35 40
VA [m3/h]
0
300 Fig. 8. The final charging time versus air volume flow rate for three values
of PCM temperature changes.
250
VA,IN = 40 m3/h
200

150 85
TPCM,0 -TPCM, ∞ = 39 ºC
100 75
65
50 55
45
0
0 25 50 75 100 125 95
Simulation
85
t [min] Experiment
75
t [min]

Fig. 7. The heat flux emitted by the PCM as a function of time. 65

55 TPCM,0 -TPCM, ∞ = 44 ºC
45
95
can be observed for the final phase change time between the 85
smallest and largest capsule size. For the charging and dis- 75
65
charging cycles, the maximum difference is equal to 31%
55 TPCM,0 -TPCM, ∞ = 49 ºC
and 33%, respectively.
45
If we decrease the volume of the capsule (beneath 20 25 30 35 40
100%), the trends of the curves shown in Figs. 12 and 13 VA [m3/h]
are approximately the same, but a great difference is
Fig. 9. The final discharging time versus air volume flow rate for three
observed when we increase the size. In this case, the char- values of PCM temperature changes.
acteristics become more flat and the heat flux out of the
analyzed device becomes more uniform in time. Smaller
dimensions of the PCM capsules, for which the volume is
equal to or less than the size of the tested polyethylene film Table 3
bags, are recommended for heating applications. Parameters of the TES unit
The computer program TES_MZ can be applied as a Parameter Value
useful tool for designing HVAC systems that cooperate
Charging temperature (C) 70
with short term TES devices. An exemplary application Charging/discharging air flow rate (m3/h) 300
of this code is presented below. Discharging temperature (C) 20
The proposed TES device can work as a part of a forced Mass of PCM (kg) 100
warm air heating system. Fig. 14 shows a schematic of this Ambient air temperature for TES units (C) 15
Thermal resistance of devices insulation (m2 K/W) 2.5
kind of installation.
162 M. Zukowski / Energy Conversion and Management 48 (2007) 155–165

450 4000
capsules volume
10, 30, 100, 300, 800%
400
3000
Charging

QPCM [W]
350
t [min]

Discharging
2000
300

1000
250

200 0
0 0.005 0.01 0.015 0. 02 0.025
0 100 200 300 400
d [m]
t [min]
Fig. 10. The time of the phase change process as a function of the distance
Fig. 12. The heat flux absorbed by the PCM as a function of time and
between PCM capsules.
capsules volume.

500 4000
capsules volume
450 10, 30, 100, 300, 800%
3000
400 Charging
QPCM [W]
t [min]

Discharging 2000
350

300
1000
250

200 0
0 100 200 300 400 500 600 700 800 0 50 100 150 200 250 300
Capsule volume [%] t [min]

Fig. 11. The time of phase change process as a function of the volume of Fig. 13. The heat flux emitted by the PCM as a function of time and
the PCM capsules. capsules volume.

Fig. 14. Concept for the short term heat storage heating system.
 M. Zukowski / Energy Conversion and Management 48 (2007) 155–165 163

Fig. 15. Parallel connection of two TES units.

Table 4 ing needs. The thermal energy DEPCM that can be stored in
Operation parameters of the heating system the encapsulated paraffin wax RII-56 is a function of the
Parameter Value mass of the PCM and its range of temperature change.
Charging temperature (C) 70 In order to design the TES device correctly, first we have
Charging air flow rate (m3/h) 300 to estimate the amount of the required storage energy
Discharging temperature (C) 20 (Eq. (19)) and the maximum supply air temperature
Discharging air flow rate (m3/h) 200
Mass of PCM (kg) 2 · 133.34 DEPCM ¼ QD dtI 3600: ð19Þ
Ambient air temperature for TES units (C) 15
Thermal resistance of devices insulation (m2 K/W) 2.5
We can obtain the mass of PCM required from Eq. (20) in
the next step but only for the complete phase change

To secure a more uniform rate of heat leaving the TES

system, we must use two devices. As Fig. 15 shows, they
can be connected in parallel with each other.
In the present study, 24 h of operation of the heat stor- 3000
age devices is analyzed. Table 4 presents the conditions that 2000
are taken into account in the simulations. 1000
It is considered that the air is warmed in an electric fur- 0
nace. The double electricity tariff is used. For example, in -1000 0 2 4 6 8 10 12 14 16 18 20 22 24
Poland, the night tariff (2200–600 and 1300–1500) is 50% of -2000
the day tariff. The consumption of electricity at off peak -3000 I TES unit
hours provides an economic benefit, and electric heating -4000
may be competitive with another fuel option. However, 3000
in order to use periodic heating, it is necessary to employ 2000 II TES unit
the TES system. It is considered that the second storage
1000
QPCM [W]

unit starts its discharge cycle when the outlet temperature

from the first unit drops below 40 C. Fig. 16 shows the cal- 0
culation results. The heat flux emitted and absorbed by the -1000
0 2 4 6 8 10 12 14 16 18 20 22 24

TES devices is presented as a function of time.

The total energy that can be received from the analyzed -2000
4000
system during the day time is about 99,300 kJ. As we can
see in Fig. 16, it is hard to get uniform heat flux out of 2000
the tested devices. Therefore, it will be necessary to 0
mix the return air from the heating space with warm air 0 2 4 6 8 10 12 14 16 18 20 22 24
from the TES units in order to stabilize the temperature -2000
of the supply air at an approximately constant level. -4000 I TES unit+II TES unit
These types of dynamic simulations are necessary to
-6000
determine properly the required PCM mass and support
to design automatic control systems serving building heat- Fig. 16. Twenty four hours operating schedule of heat storage devices.
164 M. Zukowski / Energy Conversion and Management 48 (2007) 155–165

DEPCM order to secure a more uniform heating power supply, it

mPCM ¼ ; ð20Þ
2:17DT PCM þ 150 is necessary to connect some storage devices in parallel
and to apply a discharging schedule. An analytical correla-
where
tion, Eq. (20), which includes a wide range of design
DT PCM ¼ jT PCM;0  T PCM;1 j: ð21Þ parameters, for calculation of the required PCM mass
The last step includes establishing a schedule for the dis- was established.
charge period and the number of TES devices based on From numerical analyses of the influence of the capsules
the computer simulations. geometry on the thermal performance of the TES device, it
Of course, this method is a simplified way of calcula- was concluded that the final charging and discharging time
tions. For precise sizing of the TES system, it is necessary depends exponentially on the capsules density, but the
to include the energy that is stored in the building materials maximum difference is quite small and is less than 15–
and furniture. In this case, the characteristics, presented in 18%. Moreover, smaller dimensions of the PCM capsules,
Fig. 16, can be input data for professional software tool for for which the volume is equal to or less than the size of
building simulation. This problem was widely discussed in the tested polyethylene film bags, are recommended for
Ref. [19]. heating applications.

6. Summary and conclusions Acknowledgements

The current study presents the application of the encap- This investigation was performed under grants of the
sulated paraffin wax RII-56. The mathematical model and Polish State Committee for Scientific Research No.
its numerical approximation were created to assess the 4T10B01324 and Bialystok Technical University W/IIS/
thermal performance of the latent heat storage unit. 36/03.
Three-dimensional transient simulations were applied to
the heat transfer by forced air convection, conduction
and solid–liquid phase change in the ventilated TES device. References
The effective heat capacity method was adapted to describe
the solidification and melting process, which occurs in non- [1] Zalba B, Marin JM, Cabeza LF, Mehling H. Review on thermal
isothermal conditions. The author proposed a new energy storage with phase change: materials, heat transfer analysis
and applications. Appl Therm Eng 2003;23:251–83.
approach for approximating the specific heat of the PCM [2] Khudhair AM, Farid MM. A review on energy conservation in
as a function of its temperature for all the range of operat- building applications with thermal storage by latent heat using phase
ing conditions. The interpolating cubic spline function change materials. Energ Convers Manage 2004;45:263–75.
method was used for determining the value of cp,P in each [3] Watanabe T, Kanzawa A. Second law optimization of a latent heat
time step of the temperature calculations. This type of storage system with PCMs having different melting points. Heat
Recov Syst CHP 1995;15(7):641–53.
approximation enables simulating each phase change prob- [4] Watanabe T, Kikuchi H, Kanzawa A. Enhancement of charging and
lem in the most precise way. Heat and mass transfer equa- discharging rates in latent heat storage system by use of PCM with
tions with associated boundary and initial conditions were different melting temperature. Heat Recov Syst CHP 1993;13:
solved by the control volume finite difference method with 57–66.
fully implicit scheme. The model developed in this work is [5] Arkar C, Medved S. Enhanced solar assisted building ventilation
system using sphere encapsulated PCM thermal heat storage. In: IEA,
characterized by a particular level of precision by consider- ECESIA, Advanced thermal energy storage through phase change
ation of heat losses through the insulation, three-dimen- materials and chemical reactions—feasibility studies and demonstra-
sional transient analysis of the phase change process and tion projects, 2nd workshop, Ljubljana; 2002.
heat capacity of the fluid and the support construction. [6] Ismail KAR, Henriquez JR. Numerical and experimental study of
The accuracy of the proposed model was reported too. spherical capsules packed bed latent heat storage system. Appl Therm
Eng 2002;22:1705–16.
The agreement between physical reality and the numerical [7] Pluta Z. Numerical analysis of pebble/PCM bed storage systems for
simulations was reasonable for the melting as well as for solar air heating installations. In: Proceedings of the 9th international
the solidification process. The relative maximum error in conference on thermal energy storage—The Futurestock’2003, War-
the heat flux absorbed by the paraffin wax was less then saw; 2003.
12%, and for the discharging process, the relative maxi- [8] Zukowski M. Heat transfer and pressure drop characteristics of the
underfloor air distribution system. Energ Build 2005;37(8):
mum error in the heat flux emitted was less than 10%. 890–6.
The analysis showed that the average charge time of the [9] Zukowski M. A numerical model of the air-underfloor heat
tested units was over twice higher than the time of the dis- exchanger. In: Proceedings of the 9th international conference on
charge cycle. Therefore, it is recommended to use a variable thermal energy storage—The Futurestock’2003, Warsaw;
or two speed fan control to increase the charge air flow 2003.
[10] Zukowski M. An experimental study of a short term thermal energy
rate. storage unit based on an enclosed phase change material in
As the calculation results showed, the rate of heat leav- polyethylene film bag. Energ Convers Manage, in press, doi:10.
ing the storage system strongly depends on the time. In 1016/j.enconman.2006.04.020.
 M. Zukowski / Energy Conversion and Management 48 (2007) 155–165 165

[11] Schumaker LL. Spline functions: basic theory. New York: John [16] Laouadi A, Lacroix M. Thermal performance of a latent heat energy
Wiley & Sons Inc; 1981. storage ventilated panel for electric load management. Int J Heat
[12] Atkinson K. Elementary numerical analysis. New York/Chichester/ Mass Transf 1999;24:275–86.
Brisbane/Toronto/Singapore: John Wiley& Sons, Inc; 1993. [17] Silva PD, Goncalves LC, Pires L. Transient behaviour of a latent-heat
[13] www.eti.pg.gda.pl/~tadra/tadra_pliki/SplineInter.doc [27.10.2005]. thermal energy store: numerical and experimental studies. Appl Energ
[14] Farid M, Kanzawa A. Thermal performance of a heat storage module 2002;73:83–98.
using PCMs with different melting temperatures. J Solar Energ Eng [18] Patankar SV. Numerical heat transfer and fluid flow. Series in
1989;111(2):152–7. computational methods in mechanics and thermal sciences. New
[15] Hirata T, Nishida K. An analysis of heat transfer using equivalent York: McGraw-Hill; 1980.
thermal conductivity of the liquid phase during melting inside an [19] Dumortier D, Kammerud R, Birdstall B, Andersson B, Eto J, Carroll
isothermally heated horizontal cylinder. Int J Heat Mass Transf WL, et al. Thermal energy storage system sizing. In: Proceedings of
1989;32:1663–70. building simulation’89, Canada; 1989.